NCJ Number
220643
Journal
Law Enforcement Technology Volume: 34 Issue: 10 Dated: October 2007 Pages: 134,136,140
Date Published
October 2007
Length
6 pages
Annotation
This article describes how advances in computer software, combined with state-of-the art audio equipment and experienced audio analysts, have improved the ability of investigators to detect and identify individuals from their voice patterns.
Abstract
A number of companies produce forensic audio analysis software packages. These vary in the types and depth of analysis they perform as well as in price, ranging from $400 to $1,500. Among the companies making this software are enhanced Audio, Inc., of York, PA; Cedar Audio of Portland, ME; Clarifying Technologies of Raleigh, NC; and Sound Forge, a division of Sony Corporation. This software, however, cannot compensate for the poor recording quality of the device that captures the voice. Although a recording may sound clear to the average listener, the level of background noise, the ability to reliably reproduce speech frequencies without distortion, and the audio frequency spectrum a device can record are factors that influence the quality of an audio recording for evidentiary purposes. The handling of audio recordings and tapes once they have been seized as evidence is also important in preserving their existing quality. Audio evidence should be immediately placed in an evidence bag or appropriate container and shipped to the audio lab. Also, in undercover work, the quality of recordings is influenced by the performance of the recording device, the time set for recording length (the longer the recording, the lower the audio quality), and the location of the recorder relative to the source of background noise. Forensic audio analysis software can salvage some audio evidence acquired under poor conditions, but analysis is more difficult. In addition, the analysis of audio evidence, even with the use of software, depends on the knowledge and experience of the analyst. A well-trained technician is essential especially in the area of speech recognition.